Aluminum chloride as a solid is not a strong Lewis acid

8314

J. Phys. Chem. B 2006, 110, 8314-8319

Aluminum Chloride as a Solid Is Not a Strong Lewis Acid J. Krishna Murthy,† Udo Gross,† Stephan Ru1 diger,† V. Venkat Rao,‡ V. Vijaya Kumar,‡ A. Wander,§ C. L. Bailey,§ N. M. Harrison,§,⊥ and Erhard Kemnitz*,† Institute of Chemistry, Humboldt- UniVersity Berlin, Brook-Taylor-Str. 2, 12489 Berlin, Germany, Indian Institute of Chemical Technology, Hyderabad, India, CCLRC Daresbury Laboratory, Warrington, Cheshire, WA4 4AD, UK, and Department of Chemistry, Imperial College London, Exhibition Road, London, SW7 2AZ ReceiVed: January 9, 2006; In Final Form: March 9, 2006

Aluminum chloride is used extensively as Lewis acid catalyst in a variety of industrial processes, including Friedel-Crafts and Cl/F exchange reactions. There is a common misconception that pure AlCl3 is itself a Lewis acid. In the current study, we use experimental and computational methods to investigate the surface structure and catalytic properties of solid AlCl3. The catalytic activity of AlCl3 for two halide isomerization reactions is studied and compared with different AlF3 phases. It is shown that pure solid AlCl3 does not catalyze these reactions. The (001) surface of crystalline AlCl3 is the natural cleavage plane and its structure is predicted via first principles calculations. The chlorine ions in the outermost layer of the material mask the Al3+ ions from the external gas phase. Hence, the experimentally found catalytic properties of pure solid AlCl3 are supported by the predicted surface structure of AlCl3.

Introduction AlCl3 is commonly used as a catalyst in many industrial processes. Together these processes account for more than 40% of the AlCl3 produced annually.1 These processes are all examples of Friedel-Crafts catalytic reactions catalyzed by strong Lewis acids.2 This has led to a common assumption that pure AlCl3 is a strong Lewis acid. Principally, AlCl3 and AlF3 phases should behave quite similarly concerning their Lewis acidity3 and their Lewis aciddependent catalytic performance. However, generally AlF3 is regarded as a weak Lewis acid and a poor catalyst because of the experiments performed with the crystalline R- and β-AlF3 phases. Actually, large differences between the different phases of AlF3 are observed,4 which can be explained by the substantial differences in the accessibility of the aluminum sites at the surface. The surface of well crystallized R-AlF3 shows no Lewis acidity or catalytic activity,4 and theoretical calculations have shown that the Al atoms are effectively covered by fluorine atoms.5 Conversely, the structure of beta-AlF3, which shows moderate Lewis acidity and catalytic activity,4 has been shown to contain under-coordinated 5-fold aluminum ions at the surface,6 thus, explaining the observed moderate surface acidity and catalytic activity of this phase. Just like R-AlF3, solid AlCl3 is a highly ordered crystalline material.7 Therefore, catalytic inertness should also be expected for the solid, highly crystalline phase. This is in contrast to traditional opinion based on its widespread use as catalyst. To gain insight into the Lewis acid properties of solid AlCl3 its catalytic activity for reactions, which are known to proceed exclusively when catalyzed by Lewis acids, has been investi* Corresponding author fax: (+49) [email protected]. † Humboldt- University Berlin. ‡ Indian Institute of Chemical Technology. § CCLRC Daresbury Laboratory. ⊥ Imperial College London.

30-2093-7277.

E-mail:

gated experimentally and compared with different AlF3 phases, which differ markedly in their respective degree of order. That means, a high degree of disorder is the prerequisite of undercoordinated surface Al-sites, and consequently, differently accessible surface Lewis-acid sites resulting in different catalytic properties of the chemically equally composed phases. The composition and structure of the natural cleavage plane, the (001) surface of AlCl3, has also been predicted using first principles calculations The details of the experimental work are first discussed followed by the main results of the experiments. The theoretical methodology is then described and it is followed by its main finding. The findings of the two studies are then compared and the agreement between the results is discussed. Experimental Methodology The catalytic test reactions selected are specifically Lewis acid catalyzed and truly heterogeneous, as opposed to reactions where AlCl3 becomes dissolved in the reaction mixture, e.g., in Friedel-Crafts reactions. The reactions used (viz. reaction 1 and 2) are isomerizations of fully halogenated compounds.

CCl2FCClF2 f CCl3CF3

(Reaction 1)

CBrF2CBrFCF3 f CF3CBr2CF3 (Reaction 2) Reaction 1 is of economic interest, CCl3CF3 is a sought after resource for the increasingly important CF3 chemistry.8,9 The reaction can be performed under defined heterogeneous solid/ gaseous and solid/liquid conditions, since CCl2FCClF2 is a very poor solvent for inorganic compounds. The equation for reaction 1 given above does not fully represent the possible reaction path, because consecutive dismutation reactions of CCl3CF3 may take place depending on the catalyst and the temperature needed for this reaction. The isomerization reaction of CCl2FCClF2 was shown to occur via an intramolecular reaction path, whereas

10.1021/jp0601419 CCC: $33.50 © 2006 American Chemical Society Published on Web 04/04/2006

Aluminum Chloride as a Solid Is Not a Strong Lewis Acid

J. Phys. Chem. B, Vol. 110, No. 16, 2006 8315

TABLE 1: Isomerization of CCl2FCClF2 (Reaction 1) and CBrF2CBrFCF3 (Reaction 2) over AlCl3 and Different Al-fluoride Catalysts reaction 1 liquid phase 53°C catalyst AlCl3 HS-AlF3 ACF R-AlF3 β-AlF3 a

Xa

liquid phase 53°C b

a

gas phase b

a

gas phase b

a

reaction 2 b

time [min]

[%]

Y [%]

time [min]

X [%]

Y [%]

temp [°C]

X [%]

Y [%]

temp [°C]

X [%]

Y [%]

Xa [%]

60 45 2

9 7 10c

5 7 9c

180 300 20 300 300

92 24 99c 0 0

71 22 94c

50 50 50

1.5 82 100

1.4 59 96

100 100 100 300 300

1.6 94 100 0 0

0.4 83 96

0 ≈100 ≈100 0 0

Conversion. b Yield of CF3CCl3. c At 25°C.

the dismutation reactions of the C2Cl6-xFx analogues occur rather via an intermolecular reaction path.10 However, for both types of reactions, surface Lewis acid sites of crucial strength are necessary. The weaker the surface sites the higher the reaction temperature needed and, consequently, the less selective the reaction is. Reaction 2 was selected because it proceeds only under the catalytic influence of the strongest Lewis acids such as SbF5.11 AlCl3 and, for comparison, 4 aluminum fluorides, R-AlF3, β-AlF3, HS-AlF3 (high surface AlF312), and ACF (aluminum chlorofluoride13), were tested as catalysts in reaction 1 both under solid/liquid and solid/vapor phase conditions, and in reaction 2 under solid/liquid conditions. Temperature programmed desorption of adsorbed NH3 (TPD) was attempted. Experimental. AlCl3 (Sigma-Aldrich) was freshly sublimed in a Cl2-stream and stored in a drybox. The BET surface area was not detected because a totally inert sample transfer to the BET apparatus could not be realized and contact to moisture of air would change the surface. HS-AlF3 was prepared as described elsewhere.12 It had a specific surface area (SA-BET/ N2) of 210 m2/g. ACF was prepared as described in detail elsewhere.13 The product correlated with the formula AlCl0.13F2.87 (SA-BET/N2 101 m2/g). β-AlF3 was prepared according to ref 14. Phase purity was confirmed by XRD (SA-BET/N2 26 m2/g). R-AlF3 (ACROS) was used as supplied (SA-BET/ N2 13m2/g). Catalytic Reactions. CCl2FCClF2 Isomerization, Vapor Phase. The reaction was carried out passing a mixture of CCl2FCClF vapor and N2 through a bed of 2.8 mL catalyst kept in a metal tube reactor (40 cm × 0.5 cm), with a contact time of 3.7 s. The products were trapped in cold CDCl3 and analyzed by 19F NMR. Liquid Phase. About 5 mL of CCl2FCClF2 was stirred with ∼500 mg of the respective catalyst (corresponding to a molar ratio of about 7:1) at 53 °C (AlCl3, β-AlF3, HS-AlF3) or at 25 °C (ACF) for 4-5 h in a glass bulb equipped with a reflux condenser. The reaction mixture was subjected to 19F NMR analysis at regular intervals. CBrF2CFBrCF3 Isomerization. About 0.1 g of the catalyst was added to 1 mL CBrF2CBrFCF3, and the mixture was shaken for 2 h at room temperature; after the catalyst has settled down, the supernatant liquid was analyzed by 19F NMR. Temperature Programmed Desorption of Ammonia (NH3TPD). The sample (about 0.2 g) was first heated under nitrogen up to 300 °C, then at 120 °C exposed to NH3. After flushing the excess NH3 at 120 °C with N2 for 1 h and cooling to 80 °C the TPD program was started (10°/min up to 500 °C, keeping for 30 min). Desorbed NH3 was monitored continuously via IR spectroscopy (FT-IR System 2000, Perkin-Elmer).

Experimental Results The hypothesis of this work is that, in analogy to perfectly crystalline R-AlF3, a perfectly crystalline AlCl3-surface should not exhibit any under-coordinated surface Al-site, that means it should not provide Lewis acidic surface sites and, consequently, should not be catalytically active as long as no change of the surface constitution/morphology takes place. Therefore, to prevent any chemical impact on the AlCl3 surface, catalytic reactions were selected, which (i) are exclusively catalyzed by Lewis acid sites, and (ii) the partners of which can neither AlCl3 nor AlF3 dissolve. Additionally, reaction 1 was carried out under gas-phase conditions making dissolution of the catalyst impossible. AlCl3 has previously been used for reaction 1.15 However, neither in liquid CCl2FCClF2 at room temperature over 3 h nor with CCl2FCClF2 vapor up to 100 °C did AlCl3 give any indication of catalytic activity. In a separate run, the amount of AlCl3 was enlarged to such an extend that under assumption of a BET/N2-surface area of about 10 m2/g the same absolute surface area was provided. Even under these conditions, no conversion was observed. But upon repeating the reaction with refluxing CCl2FCClF2, the isomerization started slowly (Table 1), reaching 3% conversion after 30 min, 9% after 1 h, 92% after 3 h, and completeness was obtained in 4 h. In reaction 2, AlCl3 did not show any catalytic activity. Both R-AlF3 and β-AlF3, the latter reported16,17 to be an active halogen exchange catalyst, did not show any catalytic activity either for reaction 1 up to 300 °C in the gas-phase reaction (with β-AlF3, the reaction is reported to start above 327 °C,14 crystallized R-AlF3 did not show catalytic activity under either condition, even not using a large excess of catalyst) or for reaction 2 (Table 1). Freshly prepared HS-AlF3 as well as ACF were active catalysts under both liquid and vapor phase conditions for both reactions 1 and 2 (Table 1). In the vapor phase of reaction 1 the catalytic activity of HS-AlF3 and ACF was almost comparable at 100-200 °C, but at 50 °C HS-AlF3 needed more extended contact time than ACF to reach 100% conversion. However, the selectivity of HS-AlF3 was lower than that of ACF, and with decreasing contact time, the activity was more drastically decreased than with ACF. In the liquid phase of reaction 1, HS-AlF3 was not as active as ACF, after 5 h of reaction time the conversion had reached only 24%; however, in reaction 2 both aluminum fluorides were equally highly active. Temperature programmed desorption of adsorbed NH3 (TPD) provides information about strength and quantitative distribution of acid sites on solids. Attempted TPD was not successful with AlCl3 because of its too-ready sublimation (starting already at ca. 120 °C), whereas, in case of the aluminum fluorides, it was

8316 J. Phys. Chem. B, Vol. 110, No. 16, 2006

Murthy et al. TABLE 2: Optimized Bulk AlCl3 Structurea parameter

this study

experimental

a b c B

6.053 Å 10.500 Å 6.160 Å 107.04°

5.914 Å 10.234 Å 6.148 108.25°

atom

x(fractional)

y (fractional)

z(fractional)

Al Cl1 Cl1

0.0 0.2069(0.2147) 0.2481 (0.2482)

0.1660 (0.1662) 0.0 0.1810 (0.1787)

0.0 0.2220 (0.2263) 0.2249 (0.2248)

a

Values shown in italics are constrained by summetry and were not optimized.

Figure 1. NH3-TPD of ACF, HS-AlF3, and β-AlF3. (Note that ACF decomposes above 400 °C giving rise to additional NH3 liberation, whereas HS-AlF3 is thermally stable up to 560 °C.)

successful. The TPD curves for the fluorides are shown in Figure 1 indicating superior Lewis acidity of ACF and HS-AlF3. Theoretical Methodology The relative stability of a surface with variable stoichiometry is determined by minimization of the surface free energy. The effect of an external atmosphere of chlorine on the surface structure of AlCl3 is included in the first priciples calculation via the inclusion of a chlorine chemical potential.18-20 The same methodology has been used by us in studies of AlF35,6 Calculations were performed using a linear combination of atomic orbitals scheme as implemented in the CRYSTAL code.21 The B3LYP hybrid exchange functional, which has been shown to provide reliable structures and energetics in a wide range of materials,22 was employed to approximate electronic exchange and correlation. Local Gaussian basis sets for the Al3+ and Cl- ions where obtained from standard sources.23 The relative charges on the atoms were calculated using a mulliken partition of the total charge density. This is a somewhat arbitrary choice, since there is no unique method of performing the partition of the charge density. However, the choice of a given scheme is still extremely useful in comparing the results of calculations performed using similar basis sets.24 Therefore, it provides a useful tool for comparing charge distributions of bulk and surface calculations, and for examining the effects of differing treatments of electronic exchange and correlation. The bulk unit cell of AlCl3 is defined by 10 parameters. These consist of four lattice parameters and six fractional coordinates to define the atom positions. The lattice parameters are the a, b, and c lattice vectors of the unit cell and the angle between the b lattice vector and the ac plane. There are five fractional atom coordinates required for the chlorine ions and one parameter required for the aluminum ions. We began our investigation by performing a full structure optimization of the bulk geometry of AlCl3. Optimization was performed by energy minimization using an unconstrained Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm as implemented in the DOMIN software.25 The atomic positions were allowed to relax in all directions consistent with the symmetry. To optimize the bulk cell unit of AlCl3 the stresses arising from changes in lattice parameters were calculated by numerical differentiation with a finite difference step of 0.002 Å. The optimization of atomic positions was performed using analytic derivatives of the energy.

The results of the lattice optimization and the atom positions for the bulk AlCl3 are shown in Table 2. The calculated equilibrium lattice constants for the unit cell and the nonsymmetry fixed positions of the atoms are in close agreement with those observed from experimental study.26 The maximum deviation of any parameter is 3.6%, and the mean deviation is 1.3%. AlCl3 has a layered structure. From our knowledge of such structures it is highly likely that the (100) surface will be the natural cleavage plane and, consequently, will dominate the surface morphology of any real sample. The (001) surface of the AlCl3 was obtained by cleaving the bulk crystal to produce a two-dimensional slab of material. The boundary condition perpendicular to the slab is that the wave function should decay to zero at infinity. The optimized surface structures were obtained by energy minimization with respect to the atomic positions using analytical first derivatives and the same BFGS algorithm that was used to calculate the optimized bulk structures. The geometry of each of the structures was assumed to have converged when the residual forces along all allowed symmetry directions were below 5 × 10-4 Hartrees/ Bohr. The initial termination considered consisted of a slab terminated by a complete layer of chlorine atoms. Due to the stability of the individual layers it would seem probable that this will be the most stable surface structure available. The optimization was repeated for increasing slab thickness. The geometry and surface energy were found to have converged to better than 0.01 Å and 0.001 Hartrees for a nine layer slab, e.g., three layers of the ClAlCl repeat unit. This was then used for all subsequent calculations. Three further surface structures of AlCl3 were considered. These are obtained by successive removal of chlorines. These surfaces are terminated by a layer containing two Cl ions, one Cl ion, and an Al ion, and are denoted as 2Cl, 1Cl, and Al, respectively. The geometry of each of these slabs was also fully relaxed and their relative energies were evaluated as a function of external chlorine chemical potential.18-20 Theoretical Results The free energies as a function of external partial pressure of chlorine are plotted in Figure 2 for each of the four terminations. It can be seen that the termination involving a complete layer of Cl ions is the most stable at all chlorine chemical potentials. The surface energy of this termination is 0.030 Jm-2. This is likely to be an underestimate of the surface energy as the individual layers of AlCl3 interact only via van der Waals type interactions which are underestimated by density functional calculation. This is a very low value for a surface energy and indicates that real samples will be dominated by the (001) surface.

Aluminum Chloride as a Solid Is Not a Strong Lewis Acid

J. Phys. Chem. B, Vol. 110, No. 16, 2006 8317

Figure 2. Dependence of surface energy from Cl chemical potential for differently terminated AlCl3 surfaces (terminated by Al atoms; or by Al covered by 1, 2, or 3 chlorine atoms, respectively).

Figure 3. Optimized chlorine terminated AlCl3 slab.

The surface has undergone only a very small reconstruction from the bulk cleaved surface. The largest single atom displacement is 0.0334 Å. The optimized structure is shown in Figure 3. It can be seen that the surface layer is structurally very similar to that of the bulk. The mulliken analysis shows that there is very little difference in the ion charges between the bulk terminated and the relaxed surfaces. The charge on the bulk Al is 11.092 compared to 11.109 on the Al ions in the surface layer. The charge on the bulk Cl ions is 17.640 and 17.628, the charge on the Cl ions in the surface layer is 17.601 and 17.603. Within Lewis octet theory, a Lewis acid is defined as an electron pair acceptor species. A common measure of Lewis acidity is to measure the adsorption energies of electron donating species such as ammonia onto the surface of the AlCl3. High absorption energies indicate strong Lewis acidity. Lewis acids all have a vacant orbital and/or an available LUMO and all species with full or partial positive charge behave as Lewis acids. Therefore, an alternate approach that should allow the characterization of possible acid sites on the surface

Figure 4. “Exemplary” Lewis acid AlCl3 is not a strong Lewis acid in the solid state! Thus, it resembles R- and β-AlF3 as proven by catalytic experiments. Responsible are halogen atoms forming the outer surface and covering the aluminum sites as shown by surface energy calculations.

is to consider the electrostatic potential on a plane 1 Å above the surface. A large positive potential would indicate an electron acceptor species and, hence, Lewis acidity. It is found that there is a relatively weak negative potential above the chlorines and weak positive potential immediately above the Al3+ ions, this is shown in Figure 4. Comparison The predictions discussed concerning the Lewis acidity of crystalline AlCl3, based on our theoretical calculations, are fully confirmed by the results of the catalytic test reactions 1 and 2. The results from the catalytic reactions also support previous work on the theoretical calculations for the alpha-AlF3 and betaAlF3 surfaces.5,6 In solid AlCl3, the surface Al sites are completely covered by chlorine atoms in its energetically favored state, it behaved like R-AlF3: it did not exhibit catalytic activity to be attributed to Lewis acidity. The slowly starting isomerization with

8318 J. Phys. Chem. B, Vol. 110, No. 16, 2006 refluxing CCl2FCClF2 cannot be seen as counter-evidence; on the contrary, it was the result of a heterogeneous fluorination reaction between CCl2FCClF2 and AlCl3 resulting in partial formation of ACF, the real catalytically active agent. The elemental analysis proved that AlCl3 is partially fluorinated (18.6% F; 57.4% Cl; AlCl3 requires 79.75% Cl). It is known that AlCl3 is converted into ACF by CCl3F in a highly exothermic halogen exchange reaction, whereas CCl2FCClF2 is much less reactive and needs extended heating.27 Studying reaction 1 with [36Cl] labeled AlCl3, Winfield et al.28 reported that adsorption of CCl2FCClF2 at AlCl3 sites results exclusively in chlorination of the CCl2FCClF2, while isomerization occurs exclusively at AlCl2F sites. Hence, it is concluded that isomerization occurs only at a surface of aluminum (III) in a disordered environment, rather than at the regularly ordered crystalline AlCl3 surface. Therefore, for the reaction with CCl2FCClF2, AlCl3 can be seen as precursor rather than a catalyst. Upon contact with CCl2FCClF2, vapor, up to 125 °C, fluorination of AlCl3 to form ACF did not occur, this is in line with the behavior of CCl3F which reacts with AlCl3 as a liquid but not as a gas,29 consequently almost no isomerization of CCl2FCClF2 was detectable, even when a large excess of AlCl3 was used. Crystalline R-AlF3, where the surface Al sites are almost totally blocked by fluorine atoms, and β-AlF3, the Lewis surface sites of which are too weak, were also not catalytically active. In contrast, HS-AlF312 and ACF13 are highly disordered so that Al sites are accessible resulting in the predicted very high Lewis acidity3 and, consequently, very high catalytic activity. ACF, being the more distorted material, exhibited a somewhat higher activity than did HS-AlF3 in reaction 1 only. The high selectivity toward CCl3CF3, observed for reaction 1, in both cases is mainly due to the low temperature needed because of the high Lewis acidic strength of surface sites in ACF and HSAlF3. The observation that reaction 2 proceeded quantitatively only under the catalytic influence of HS-AlF3 and ACF provides evidence that Al-F and Al-F-Cl compounds can principally exhibit Lewis acidity about as high as SbF5 in agreement with calculated strength.3 However, solid AlCl3 was similar to R-AlF3 and β-AlF3, it was not active in reaction 2; these solids are far from being strong Lewis acids, as long as they exist in highly crystalline phases, due to the type of surface termination by halogens. In solvents AlCl3 behaves quite differently: the aluminum is tetrahedrally coordinated forming quasi-tetrahedral complexes of the type AlCl3‚L or, depending on the electron donor ability of the solvent, even ionic equilibria including AlCl4- species are formed. The degree of electron polarizability and charge transfer, respectively, can be seen by the shift of the IR absorption bands of solvent molecules adsorbed at solid AlCl3. Not surprisingly, CHCl3, a compound whose properties are closely related to CCl3F, shows no solvent interactions.30 Although AlF3 should be about as Lewis acidic as AlCl3 (according to ab initio calculated fluoride affinity data3), the latter cannot be replaced by AlF3 in nearly all reactions catalyzed by AlCl3 because AlF3 is not soluble in the reaction mixtures used. A comparison of the differences of standard enthalpies of formation for gaseous and condensed state for AlCl3 (∆Hf ) 121 kJ/mol) and AlF3 (∆Hf ) 301 kJ/mol) and the large differences in Al-X binding energies shows the reason for the AlF3 insolubility. Only in a very highly disordered state as in ACF and in HS-AlF3 Al sites will become accessible to potential reactants resembling somewhat the dissolved state of

Murthy et al. AlCl3, so that the predicted high Lewis acidity becomes reality and the materials are highly catalytically active. Conclusion According to Christe et al.3 and our own calculations31 all aluminum halides should be strong Lewis acids. However, as long as AlCl3, like AlF3, exist in a solid crystalline state, they cannot be regarded as being Lewis acidic. The explanation for the fact that solid AlCl3 and crystalline AlF3 phases do not behave as strong Lewis acids, as expected from the earlier theoretical calculations, is that their surfaces are chlorine or fluorine terminated, respectively, giving any reactant no opportunity to coordinate Al3+ sites. This is in agreement with our surface energy calculations for solid AlCl3. If the aluminum halide structure becomes disordered, as in HS-AlF3 or ACF and most extremely upon dissolution of AlCl3, these compounds exhibit strong Lewis acidity. However, in solution, AlCl3 is no longer a solid, and hence, it is no longer a heterogeneous catalyst. Consequently, it is not a question whether the surface area of the AlX3-phase is large, but rather if coordinatively unsaturated Al-sites are accessible for a reactant or not. Well crystallized AlF3 and AlCl3 surfaces, respectively, do not exhibit under-coordinated surface Al-sites, but HS-AlF3 and ACF do. Hence, it is not surprising that increasing the amount of well crystallized AlX3 (X ) F, Cl), to provide the same absolute surface area, does not result in catalytic activity. On the other hand, if suitable synthesis procedures are applied, resulting in strongly distorted solid phases, as for ACF or HSAlF3, strong solid Lewis acids may be obtained because of coordinatively unsaturated Al-surface sites giving access for suitable reactants. Acknowledgment. We thank the German BMBF (Federal Ministry for Education and Research) for support of part of this work via DLR (German Centre of Aviation and Astronautics) Indian-German Scientific Cooperation Program. We thank the EU for support of part of this work through the 6th Framework Program (FUNFLUOS, contract no. NMP3-CT-2004-5005575). References and Notes (1) Greenwood, N. N.; Earnshaw, A. Chemistry of the Elements, Butterworth Heinmann, Cambridge UK 1984. (2) Olah, G. A. Friedel-Crafts and Related Reactions; Interscience: New York 1963; Vols 1-4, (esp. chap 1 and 2). (3) Christe, K. O.; Dixon, D. A.; McLemore, D.; Wilson, W. W.; Sheehy, J. A.; Boatz, D. A. J. Fluorine Chem. 2000, 101, 151. (4) Kemnitz, E.; Winfield, J. M. Fluoride catalysts and their application to heterogeneous catalytic fluorination related reactions. In AdVanced Inorganic Fluorides: Synthesis, Characterization and Applications; Nakajima, T., Tressaud, A., Zemva, B., Eds.; Elsevier Science: New York, 2000; pp 367-402. (5) Wander, A.; Searle, B. G.; Bailey, C. L.; Harrison, N. M. J. Phys. Chem. B 2005, 109, 22935. (6) Wander, A.; Bailey, C. L.; Searle, B. G.; Mukhopadhyay, S.; Harrison, N. M. Phys. Chem. Chem. Phys. 2005, 7, 3989. (7) Troyanov, S. I. J. Inorg. Chem. (Russ.) 1992, 37, 266. (8) Nenajdenko, V. G.; Varseev, G. N.; Shastin, A. V.; Balenkova, E. S. J. Fluorine Chem. 2005, 126, 907; and citations therein. (9) Ru¨diger, St.; Gross, U.; Chandra Shekar, S.; Venkat Rao, V.; Sateesh, M.; Kemnitz, E. Green Chem. 2002, 4, 541. (10) (a) Bozorgzadeh, H.; Kemnitz, E.; Nickkho-Amiry, M.; Skapin, T.; Winfield, J. M. J. Fluorine Chem. 2001, 107, 45. (b) Bozorgzadeh, H.; Kemnitz, E.; Nickkho-Amiry, M.; Skapin, T.; Tate, G. D.; Winfield, J. M. J. Fluorine Chem. 2001, 112, 225. (11) Petrov, V. A.; Krespan, C. G.; Smart, B. E. J. Fluorine Chem. 1998, 82, 125. (12) Kemnitz, E.; Gross, U.; Ru¨diger, St.; Shekar, C. S. Angew. Chem., Int. Ed. 2003, 42, 4251. (13) Krahl, T.; Sto¨sser, R.; Kemnitz, E.; Scholz, G.; Feist, M.; Silly, G.; Buzare, J.-Y. Inorg. Chem. 2003, 42, 6474.

Aluminum Chloride as a Solid Is Not a Strong Lewis Acid (14) Bozorgzadeh, H.; Kemnitz, E.; Nickkho-Amiry, M.; Skapin, T.; Winfield, J. M. J. Fluorine Chem. 2001, 110, 181. (15) Paleta, O.; Liska, F.; Posta, A.; Dedeck, V. Collect. Czech. Chem. Commun. 1980, 45, 104. (16) Manzer, L. E.; Rao, V. N. M. U.S. Patent 4,902,838, 1990. (17) Morikawa, S.; Samejima, S.; Yoshitake, M.; Tatematsu, S.; Tanuma, T. Japan Patent 2-115135, 1990. (18) Batyrev, I.; Alavi, A.; Finnis, M. W. Faraday Discuss. 1999, 114, 33. (19) Wang, X.-G.; Weiss, W.; Shaikhutdinov, Sh. K.; Ritter, M.; Peterson, M.; Wagner, F.; Schlo¨gl, R.; Scheffler, M. Phys. ReV. Lett. 1998, 81, 1038. (20) Reuter, K.; Scheffler, M. Phys. ReV. 2002, B65, art. no. 035406. (21) Saunders, V. R.; Dovesi, R.; Roetti, C.; Orlando, R.; ZicovichWilson, C. M.; Harrison, N. M.; Doll, K.; Civalleri, B.; Bush, I. J.; D’Arco, Ph.; Llunell, M. CRYSTAL 2003 Users’s Manual; University of Torino: Torino, 2004. (22) Muscat, J.; Wander, A.; Harrison, N. M. Chem. Phys. Lett. 2001, 342, 397.

J. Phys. Chem. B, Vol. 110, No. 16, 2006 8319 (23) Basis sets are available from http://www.crystal.unito.it/Basis_Sets/ ptable.html (24) Pisani, C.; Dovesi, R.; Roetti, C. Hartree-fock ab initio Treatment of Crystalline Systems, Lecture Notes in Chemistry: Springer-Verlag: Heidelberg, 1988; Vol. 48. (25) Spellucci, P. University of Darmstadt. DOMIN software available from http://plato.la.asu.edu/donlp2.html. (26) Troyanov, S. I. Zh. Neorganicheskoi Khimii 1992, 37, 226. (27) Sievert, A. G.; Krespan, C. G.; Weigert, F. J. (to DuPont), U.S. Patent 5,157,171. 1992. (28) McBeth, D. G.; McGeough, M. M.; Webb, G.; Winfield, J. M.; McCulloch, A.; Winterton, N. Green Chem. 2000, 2, 15. (29) Krahl, T. Ph.D. Thesis, Humboldt-Universita¨t zu Berlin (Germany), 2005. (30) Bertoluzza, A.; Bonino, G. B.; Fabbri, G.; Lorenzelli, V. J. Chim. Phys. 1966, 3, 385. (31) Krahl, Th.; Kemnitz, E. J. Fluorine Chem. 2006, in press.